Transcript Genetic Epidemiology of Parkinson`s Disease (GEO

HuGENet Network of
Networks Workshop:
GEO-PD Consortium
Demetrius M. Maraganore, MD
Professor of Neurology
Mayo Clinic College of Medicine
Rochester, MN
Edmond J. Safra Global Genetics Consortia

Michael J. Fox Foundation ($1.2 million initiative)

Five grants awarded

Tatiana Foroud

Demetrius Maraganore

Haydeh Payami

Clemens Scherzer
Lorene Nelson


Collaborative studies of a chromosome
5 PD susceptibility gene
Collaborative pooled analysis of the
SNCA REP1variant and PD
Gene-environment interaction in PD:
predicting the onset, prognosis, and
response to treatment
Gene expression in PD
Genetic and environmental factors in PD
http://www.michaeljfox.org/news/article.php?id=114
Handling non-participation

Be inclusive



Recognize participants




Shared leadership (core PIs and co-PIs, Global Site PIs and co-Is)
Authorships (multiple authors per site)
Subcontracts
Foster collegiality


Invitation of all correspondence authors of published genetic
association studies for a targeted gene and disease to participate in
a collaborative pooled analysis
Invitation of additional investigators to participate (e.g.,
correspondence authors of published genetic association studies
for other genes and the same disease)
Annual meeting of the consortium
Cope

Metaanalysis of published data, including non-participating sites
 secondary analyses
Other scientific issues

Comparison subjects



Case-only studies



Correlation of genotypes to age at onset, or to prognostic
outcomes (modifier genes)
Gene interactions
Gene-environment interactions


Siblings, unrelated controls, or both
Considerations on population stratification
Likely to require prospective study design
Globally informative SNPs

Haplotype tagging, LD mapping in diverse populations
Data flow

Participant requirements


N ≥ 100 cases, 100 controls
Minimal dataset




Sample sharing


n = 20 DNAs (200 ng each)
Willingness to share de-identified individual level data


study characteristics
clinical characteristics
genotypes
supplemental data online
Transfer of minimal dataset to statistical core



Formatted Excel spreadsheet
Data archived in SAS database
Checks for missing data, errors

query sheets to investigators
Standardization of phenotypes and genotypes

Standardization of phenotypes (formatted Excel spreadsheets)

Study characteristics




Individual level data



sources of cases: community or clinic
sources of controls: community or hospital, blood bank, spouses
diagnostic criteria (references)
cases and controls: source, age at study, gender, ethnicity, genotypes
cases only: age at onset, family history (≥1 1st degree relative)
Standardization of genotypes (DNAs for re-genotyping)

List of 20 lab ids, genotypes sent to statistical core


20 DNAs (200 ng each) sent to laboratory core


heterozygosity checks
re-genotyping blinded to original allele calling
List of new genotypes sent to statistical core



tests of reliability (if < 90% reliability, the study is excluded)
post-coding of all genotypes (with laboratory core as reference)
genotyping reports to contributing sites (reliability, HWE, post-coded
genotypes, cleaned datasets)
Other standardization issues

Exclusion of studies




Failure to provide minimal datasets, DNAs by deadlines
Genotyping reliability < 90%
Lack of HWE in controls
Statistical considerations



Tests for heterogeneity, HWE
Unadjusted analyses (missing data)
Adjusted analyses (confounders)


study, age at study, gender
Stratified analyses (genetic heterogeneity)




ethnicity
age at study
gender
family history